sql >> Database teknologi >  >> RDS >> Sqlserver

SQL Server 2008 - Avanceret søgning/sortering

Du leder efter Levenshtein distance

Her er en algoritmeimplementering til T-SQL:Beregning Levenshtein Distance i TSQL

Som funktionssignatur er som:

CREATE FUNCTION [dbo].[LEVENSHTEIN]( @s NVARCHAR(MAX), @t NVARCHAR(MAX) )
/*
Levenshtein Distance Algorithm: TSQL Implementation
by Joseph Gama

http://www.merriampark.com/ldtsql.htm

Du kan kalde det i din forespørgsel som:

Select *
from ValuesInDatabase V
order by
    [dbo].[LEVENSHTEIN](  V.ClientName, 'Alliance A' ) 

Du skal ikke forvente høj ydeevne.

Redigeret

Jeg kopierer her Joseph Gamas funktion for at bevare den, glem ikke at besøge devioblog :

CREATE FUNCTION [dbo].[LEVENSHTEIN]( @s NVARCHAR(MAX), @t NVARCHAR(MAX) )
/*
Levenshtein Distance Algorithm: TSQL Implementation
by Joseph Gama

http://www.merriampark.com/ldtsql.htm

Returns the Levenshtein Distance between strings s1 and s2.
Original developer: Michael Gilleland http://www.merriampark.com/ld.htm
Translated to TSQL by Joseph Gama

Fixed by Herbert Oppolzer / devio
as described in http://devio.wordpress.com/2010/09/07/calculating-levenshtein-distance-in-tsql
*/
RETURNS INT AS
BEGIN
  DECLARE @d NVARCHAR(MAX), @LD INT, @m INT, @n INT, @i INT, @j INT,
    @s_i NCHAR(1), @t_j NCHAR(1),@cost INT

  --Step 1
  SET @n = LEN(@s)
  SET @m = LEN(@t)
  SET @d = REPLICATE(NCHAR(0),(@n+1)*(@m+1))
  IF @n = 0
  BEGIN
    SET @LD = @m
   GOTO done
  END
  IF @m = 0
  BEGIN
    SET @LD = @n
    GOTO done
  END

  --Step 2
  SET @i = 0
  WHILE @i <= @n BEGIN
    SET @d = STUFF(@d,@i+1,1,NCHAR(@i))        --d(i, 0) = i
    SET @i = @i+1
  END

  SET @i = 0
  WHILE @i <= @m BEGIN
    SET @d = STUFF(@d,@i*(@n+1)+1,1,NCHAR(@i))    --d(0, j) = j
    SET @i = @i+1
  END

  --Step 3
  SET @i = 1
  WHILE @i <= @n BEGIN
    SET @s_i = SUBSTRING(@s,@i,1)

    --Step 4
    SET @j = 1
    WHILE @j <= @m BEGIN
      SET @t_j = SUBSTRING(@t,@j,1)
      --Step 5
      IF @s_i = @t_j
        SET @cost = 0
      ELSE
        SET @cost = 1
      --Step 6
      SET @d = STUFF(@d,@j*(@n+1)[email protected]+1,1,
        NCHAR(dbo.MIN3(
          UNICODE(SUBSTRING(@d,@j*(@n+1)[email protected]+1,1))+1,
          UNICODE(SUBSTRING(@d,(@j-1)*(@n+1)[email protected]+1,1))+1,
          UNICODE(SUBSTRING(@d,(@j-1)*(@n+1)[email protected]+1,1))[email protected])
        ))
      SET @j = @j+1
    END
    SET @i = @i+1
  END      

  --Step 7
  SET @LD = UNICODE(SUBSTRING(@d,@n*(@m+1)[email protected]+1,1))

done:
  RETURN @LD
END



  1. Hent id for hver INSERT-sætning i multi-forespørgsel

  2. SQLSTATE[42S22]:Kolonne blev ikke fundet:1054 Ukendt kolonne 'id' i 'where clause' (SQL:vælg * fra 'songs' hvor 'id' =5 limit 1)

  3. GWT-database-adgang uden RPC

  4. Sådan bruger du KNN Machine Learning Model med 2UDA – PostgreSQL og Orange (del 1)